When scientists analysed the first OSIRIS-REx samples from asteroid Bennu, they found minerals no one expected: sodium carbonates, halite, and phosphates precipitated from an evaporating brine. These minerals belong on the floor of a drying desert lake, not inside a rubble-pile asteroid half a kilometre across. They are a reminder that the most primitive objects we can hold were never truly pristine. Their parent bodies were heated, flooded, and chemically reworked within the first few million years of solar system history.

Chondrites are often described as cosmic sediment. They are aggregates of chondrules, refractory inclusions, metal, and fine-grained matrix that accreted from the protoplanetary disk. That description holds only at the moment of accretion. Internal heating began to overprint the nebular record almost immediately. In ordinary and enstatite chondrites, heat drove solid-state recrystallization. In many carbonaceous chondrites, it melted accreted ice and drove fluid–rock reactions that converted anhydrous silicates into clays.

The petrologic classification of Van Schmus and Wood captures this divergence in a single scale. Type 3 sits at its least-altered centre, values rise toward 6 with heating, and they fall toward 1 with aqueous alteration. Reading that scale correctly means understanding the processes behind it and the radioactive engine that powered them. Seen this way, the meteorite record is less a catalogue of primordial ingredients than a set of thermal and hydrological experiments run on hundreds of small worlds.

The Thermal Metamorphism Sequence

Ordinary chondrites offer the clearest view of thermal metamorphism because they span the full sequence, from unequilibrated type 3 to recrystallized type 6. The most primitive members, such as Semarkona, preserve glassy chondrule mesostasis, compositionally zoned olivine and pyroxene, fine-grained opaque matrix, and presolar grains. They also retain insoluble organic matter that would not survive sustained heating. Peak temperatures for these rocks stayed within a few hundred degrees Celsius.

Through the type 3 subtypes, the changes are subtle but measurable. Chromium diffuses out of ferroan olivine and exsolves as chromite, which makes it a sensitive indicator at the lowest grades. Thermoluminescence sensitivity increases as feldspathic glass devitrifies. Raman spectra of matrix organics record the progressive ordering of carbon. Together, these indicators allow type 3 to be subdivided from 3.00 to 3.9. The subdivision matters because the earliest heating stages erase the most information about primordial materials.

Above roughly 600°C, the transformation becomes obvious in thin section. Fe–Mg interdiffusion homogenizes olivine and low-Ca pyroxene. This is why equilibrated chondrites cluster tightly in fayalite content: roughly Fa16–20 for H, Fa22–26 for L, and Fa26–32 for LL. Monoclinic low-Ca pyroxene inverts to orthopyroxene. Chondrule mesostasis crystallizes into plagioclase, which coarsens from a few micrometres in type 4 to grains exceeding 50 micrometres in type 6. Meanwhile, chondrule boundaries blur as the matrix recrystallizes into an integrated texture.

Two-pyroxene and olivine–spinel thermometry give peak temperatures of roughly 600–700°C for type 4, rising to 800–950°C for type 6. Metallographic cooling rates and radiometric ages add the dimension of time. In the classic onion-shell model, type 6 material came from a deep, hot interior and cooled slowly, while type 4 sat closer to the surface and cooled quickly. Several H chondrites follow this pattern, with Pb–Pb, Hf–W, and Ar–Ar ages that become progressively younger with increasing petrologic type.

The model is not universal, however. Some H and L chondrites have cooling rates inconsistent with any simple burial depth. This suggests that impacts fragmented and reassembled their parent bodies while the interiors were still hot. Particles returned by Hayabusa from Itokawa make the same point from a different angle: equilibrated LL4–6 material now resides in a small rubble pile. Metamorphic grade faithfully records thermal history. Present-day location, by contrast, often records collisional history, and the two must be disentangled before reconstructing any parent body.

Takeaway

A petrologic type is a thermometer that only records its maximum. Reconstructing a parent body requires the clock and the cooling rate as well, because peak temperature alone cannot distinguish burial depth from collisional reshuffling.

Aqueous Alteration Products

Carbonaceous chondrites run the petrologic scale in the opposite direction. CM2 and CR2 chondrites retain chondrules that are partially replaced by secondary minerals. CI1 chondrites and Ryugu samples are almost entirely secondary: the rock is dominated by phyllosilicates, with few or no surviving chondrules. The defining reaction is the hydration of anhydrous silicates, metal, and sulfides. The water came from melting ice that accreted alongside the rock, likely in the colder outer reaches of the protoplanetary disk.

Phyllosilicate assemblages differ systematically between groups. CM chondrites are dominated by Fe-rich serpentine, including cronstedtite, and by tochilinite–cronstedtite intergrowths that formed from altered metal and sulfide. CI chondrites and Ryugu contain intergrown serpentine and saponite, a smectite clay, reflecting more extensive alteration under different fluid conditions. In CMs, progressive alteration can be tracked by a shift from Fe-rich toward Mg-rich serpentine as chondrule olivine and pyroxene are consumed.

Carbonates carry much of the quantitative information. Calcite dominates in CMs, while CIs and Ryugu host dolomite and breunnerite alongside calcite. Clumped isotope thermometry on CM carbonates suggests formation temperatures of roughly 20–70°C. Mn–Cr dating places carbonate precipitation around four to six million years after the formation of calcium-aluminum-rich inclusions. Framboidal and plaquette magnetite, along with pyrrhotite and pentlandite, complete the assemblage. Together they record evolving redox conditions and fluid chemistry.

Sample return missions have sharpened this picture considerably. Ryugu grains, shielded from terrestrial weathering, showed that some sulfates and ferrihydrite long reported in CI meteorites were partly artifacts of their time on Earth. One Ryugu pyrrhotite crystal preserved a fluid inclusion containing CO2-bearing water. Bennu samples went further, yielding sodium carbonates, halite, sylvite, and phosphates consistent with brine evaporation. This implies that fluids became concentrated and alkaline late in the alteration sequence.

The significance extends beyond mineralogy. Aqueous alteration created environments where organic synthesis and destruction proceeded side by side. The amino acids in Murchison, and the amino acids and all five canonical nucleobases identified in Bennu samples, emerged from or survived fluid–rock systems on small, cold bodies. Altered carbonaceous asteroids preserve some of the earliest hydrothermal systems we can sample directly. They predate any surviving terrestrial record of liquid water by hundreds of millions of years.

Takeaway

The least-heated meteorites are often the most chemically transformed. Treating 'primitive' and 'unaltered' as synonyms distorts how we read the early solar system, because water was a reagent rather than a contaminant.

Aluminum-26: The Brief Engine Behind It All

Bodies tens to hundreds of kilometres across retain little accretional heat. Long-lived radionuclides such as uranium, thorium, and potassium-40 release energy far too slowly to matter over a few million years. The dominant early heat source was aluminum-26, which decays to magnesium-26 with a half-life of about 717,000 years. Calcium-aluminum-rich inclusions record a canonical initial ratio of aluminum-26 to aluminum-27 near 5 × 10⁻⁵. Their excess magnesium-26 anchors the zero point of solar system chronology.

Aluminum-26 combines high specific heating power with rapid decay, which makes accretion time the master variable. Thermal models show a sharp threshold. Planetesimals that accreted within roughly the first million years melted and differentiated, consistent with Hf–W ages for iron meteorite cores. Bodies accreting around two million years after CAIs reached metamorphic temperatures without melting. Those accreting after roughly three to four million years could melt ice and sustain liquid water, but little more.

The meteorite record fits this framework remarkably well. Chondrules in ordinary chondrites formed roughly 1.5 to 2.5 million years after CAIs, which sets a minimum accretion age for their parent bodies. Thermal models of the H chondrite parent body, constrained by onion-shell cooling ages, point to accretion near two million years and a radius of order 100 kilometres. CM and CI parent bodies accreted later with less live aluminum-26. They generated just enough heat to precipitate carbonates at the ages Mn–Cr dating indicates.

Water complicates the thermal budget. The latent heat of melting ice, the high heat capacity of water, and possible hydrothermal convection all buffer peak temperatures. Exothermic serpentinization reactions, meanwhile, add heat of their own. Formation location mattered too. Nucleosynthetic isotope anomalies indicate that carbonaceous parent bodies accreted in a distinct outer-disk reservoir, where ice was abundant and accretion may have been delayed. Iron-60 also contributed, but its revised initial abundance appears too low to compete with aluminum-26.

Puzzles remain. Paleomagnetic evidence from Allende and other CV chondrites has been interpreted as the signature of an internal dynamo. That would imply a partially differentiated parent body, with a molten core beneath an unmelted chondritic crust. Whether aluminum-26 was distributed homogeneously across the disk is also debated, and any heterogeneity would shift inferred accretion ages. The underlying logic holds nonetheless: accretion timing relative to a rapidly decaying isotope determined whether a world melted, baked, or soaked.

Takeaway

On small bodies, when material accreted mattered as much as what it was made of. A delay of a single million years could separate a body that grew an iron core from one that became a lump of wet clay.

Seen together, metamorphism and aqueous alteration are not opposing processes. They are outcomes of the same physics operating under different initial conditions. Accretion time, ice content, and body size determined whether aluminum-26 heat recrystallized silicates, mobilized fluids, or did both, as in CV and some CO chondrites where the two overlap.

This reframes the search for primitive material. Rare type 3.00 chondrites and minimally altered carbonaceous samples are survivors of a narrow window, valuable precisely because of what nearly happened to them. It also reframes habitability: liquid water, organics, and chemical energy coexisted on small bodies millions of years before Earth had oceans.

The lesson extends to exoplanetary systems, where short-lived radionuclide inventories likely vary with stellar birth environment. A system born with less aluminum-26 might build wetter, less differentiated planetesimals, and from them, planets with very different water budgets and interiors.